Track x-kit agent executions and store trace data as JSONL files.
npx skills add https://github.com/x-mesh/x-kit --skill x-trace
Please help me install this Agent Skill. Skill: x-trace Source: https://github.com/x-mesh/x-kit/tree/main/x-trace/skills/trace Command: npx skills add https://github.com/x-mesh/x-kit --skill x-trace
x-trace solves the challenge of understanding complex, multi-agent executions in the x-kit environment, providing insights into token usage, execution costs, and session replay for enhanced observability and debugging.
Use the /xm:trace start command to begin a new tracing session for your workflow.
High-intent search queries and answers about installing and using this skill.
Track token usage and execution costs by running the /xm:trace start command to begin a tracing session. The system captures real-time performance metrics and token consumption for each agent interaction within your multi-agent workflows.
Multi-agent observability provides insights into complex executions by capturing token usage, execution times, and agent interactions. You need timeline visualization to understand the sequence and duration of events when debugging intricate multi-agent workflows.
Replay a previous multi-agent execution session using the execution replay feature to re-run past workflows. This allows you to investigate specific execution steps and identify exactly where and why failures occurred.
Yes, x-trace requires saving trace data as JSONL files. This format captures the sequential execution logs, token tracking metrics, and agent interactions needed for detailed observability and cross-session comparison.
Yes, you can compare different multi-agent execution sessions to identify deviations and changes over time. Cross-session comparison highlights variations in token usage, execution times, and agent interactions across your workflows.
The best way to debug complex agent interactions in x-kit workflows is using x-trace to visualize execution timelines and replay sessions. This provides detailed observability into sequence, duration, and token costs for each step.
rysweetAnalyze claude-trace JSONL files for session health, token usage, and error patterns.
ice-droid-99Fetch aggregated MLflow trace metrics for token usage, latency, and quality scores.
Owl-ListenerIdentify visibility requirements and capture execution traces for multi-agent workflows.
chiptoe-svgAdd a local monitoring dashboard for NanoClaw deployments.
